From 1dca0db9bba7e6c5ef1ceab60b45f3ef3c731837 Mon Sep 17 00:00:00 2001
From: MeowcaTheoRange <me@abtmtr.link>
Date: Thu, 22 Feb 2024 05:29:31 -0600
Subject: [PATCH] Improve scraping opinionation (Thanks, The Quick Brown Fox -
 WANDERLUST)

---
 index.js     | 52 +++++++++++++++++++++++++++++++++++++++-------------
 package.json |  3 ++-
 2 files changed, 41 insertions(+), 14 deletions(-)

diff --git a/index.js b/index.js
index a3cdd42..50c57a2 100644
--- a/index.js
+++ b/index.js
@@ -2,13 +2,14 @@ require('dotenv').config();
 const { Readable, PassThrough } = require("stream");
 const { finished } = require("stream/promises");
 const { searchMusics } = require('fix-esm').require("node-youtube-music");
-const { GetListByKeyword } = require("youtube-search-api");
+const { GetListByKeyword, GetVideoDetails } = require("youtube-search-api");
 const skewered = require("skewered");
 const { createClient } = require("fix-esm").require("webdav");
 const path = require("path");
 const { readFileSync, writeFileSync } = require("fs");
 const io = require('socket.io-client');
 const NodeID3 = require('node-id3');
+const xpath = require("xpath-html");
 
 function pathGenerator({ url, name, channelName }) {
   return `${url}-${skewered(`${name} - ${channelName}`)}.mp3`;
@@ -26,6 +27,23 @@ function writeLastSong({ name, channelName }) {
   return null;
 }
 
+async function scrapeLastFMWebsiteForVideo(trackData) {
+  const lastFMSite = await fetch(trackData.url).then(x => x.text());
+  // classes: image-overlay-playlink-link
+  // attrs: data-youtube-id
+  
+  const attr = xpath.fromPageSource(lastFMSite).findElement("//a[contains(@class, 'image-overlay-playlink-link')]/@data-youtube-id").getAttribute("data-youtube-id");
+
+  if (attr == null) return;
+
+  const youtubeVideo = await GetVideoDetails(attr);
+
+  return {
+    ...youtubeVideo,
+    id: attr
+  };
+}
+
 async function getVideo() {
   const trackData = await fetch(
     `https://${process.env.LASTFM_INSTANCE}/2.0/?method=user.getrecenttracks&user=${process.env.LASTFM_USERNAME}&api_key=${process.env.LASTFM_API_KEY}&format=json&limit=1&extended=1`
@@ -65,23 +83,31 @@ async function getVideo() {
   const musicList = await searchMusics(`${trackData.artist.name} - ${trackData.name}`);
   
   const youtubeMusicVideo = musicList.find((song) => {
-    return skewered(song.title).includes(skewered(trackData.name));
+    return skewered(song.title).includes(skewered(trackData.name))
+        && song.artists.some(artist => skewered(trackData.artist.name).includes(artist.name));
   });
 
   if (youtubeMusicVideo == null) {
-    const videoList = await GetListByKeyword(`${trackData.artist.name} - ${trackData.name}`, false, 1, [
-      {type: "video"}
-    ]);
-  
-    const musicVideo = videoList.items[0];
+    const lastFMVideo = await scrapeLastFMWebsiteForVideo(trackData);
 
-    selectedVideo = {
-      url: musicVideo.id,
-      name: musicVideo.title,
-      channelName: musicVideo.channelTitle
+    if (lastFMVideo == null) {
+      const videoList = await GetListByKeyword(`${trackData.artist.name} - ${trackData.name}`, false, 1, [
+        {type: "video"}
+      ]);
+    
+      const musicVideo = videoList.items[0];
+
+      selectedVideo = {
+        url: musicVideo.id,
+        name: musicVideo.title,
+        channelName: musicVideo.channelTitle
+      };
+    } else selectedVideo = {
+      url: lastFMVideo.id,
+      name: lastFMVideo.title,
+      channelName: lastFMVideo.channel
     };
-  }
-  else selectedVideo = {
+  } else selectedVideo = {
     url: youtubeMusicVideo.youtubeId,
     name: youtubeMusicVideo.title,
     channelName: youtubeMusicVideo.artists[0].name
diff --git a/package.json b/package.json
index 46e6002..2c57305 100644
--- a/package.json
+++ b/package.json
@@ -20,6 +20,7 @@
     "socket.io": "^4.7.4",
     "socket.io-client": "^4.7.4",
     "webdav": "^5.3.2",
+    "xpath-html": "^1.0.3",
     "youtube-search-api": "^1.2.1"
   }
-}
\ No newline at end of file
+}